PhaST is a health information system that assists clinicians? management of medications in ambulatory settings. PhaST seeks to protect outpatients taking drugs that have recognized side effect risks even when those drugs are correctly prescribed. Existing clinical strategies for monitoring risky drugs rely on frequent office visits. Such strategies might succeed for families with access to a system that can supply the requisite visits. However, for many families, the financial or organizational burden of frequent visits is high. The result is that many patients are exposed to inadequate outpatient medication safety monitoring, with those having least access to health services likely to experience the least monitoring. The PhaST system is an automated system for monitoring of medication adherence, side effects, and patient symptoms. PhaST uses research-based assessment procedures administered using interactive voice response (IVR) telephony. When a patient reports a problem with a medication on an IVR call, PhaST alerts a nurse trained to triage the problem, to counsel the patient or family, and when necessary to contact the patient?s prescribing clinician or the hospital emergency services. PhaST is built within a hospital data warehouse and integrates fully with the electronic medical record and computerized physician order-entry system, so clinicians have real-time information on the patient?s experience with the medication. Our target medication is the pediatric use of anti-depressants. The growing use of psychoactive medications in children and adolescents has greatly increased the public health relevance of medication-related adverse events, including suicidality and manic activation. To compare PhaST and usual care, we propose a randomized trial in a large urban specialty mental health system serving a primarily Medicaid population. We will enroll 800 youths receiving new prescriptions for anti-depressants. We will assess the patients for adverse events during home visits at baseline and 1, 2, and 3 months. We will compare chart-documented adverse events against adverse events as determined by an examiner blind to the patient?s randomization. We predict higher agreement between chart-documented adverse events and examiner-determined adverse events in the PhaST condition. We will also compare PhaST and usual care on measures of patient and provider satisfaction, patient outcomes, and measures of the quality of medication management such as rates of patient medication non-adherence.